RDFa: The Five Key Takeaways

RDFa should come with the warning: "Not Intended for Human Consumption!" In this case, that's a good thing. Why? RDFa is a tad ungainly for humans to parse. However, it's not for machines, and that's precisely for whom RDFa is intended. RDFa holds the promise of expressing to machines that which is readily evident to humans. In this way, RDFa is intended as a bridge between the human and data webs.

Because RDFa is intended for machines, it must be completely unobtrusive to humans. And it is. Making use of tags already present in HTML, machine-readable data is transparently introduced by RDFa. In other words, you would need to view the source of the HTML to actually expose the RDFa for human consumption. Machines see it automagically. Most of the emphasis these days is on working RDFa into XHTML. In principle, however, RDFa can be tweaked to work with any markup language.

The Web's a wonderful place already. However, to move beyond the human-human and human-machine interactions that characterize today's Web, there is a need to facilitate machine-machine interactions. These interactions are key in enabling the meaning-centric Web envisaged originally by Web inventor Tim Berners-Lee. In and of itself, RDFa does not deliver us The Semantic Web, but it does bring realization of the same a whole lot closer.

How? RDF (Resource Description Format) representations can be readily extracted from Web pages marked up with RDFa. From these extractions, subject-predicate-object relationships can be built. It turns out that relationships are extremely important in realizing the Semantic Web. Why? When subject-predicate-object relationships are evident, connections are made between otherwise isolated snippets of data. And ultimately, this information can be further refined into knowledge. This systematic introduction of meaning is key to enabling and then leveraging a Semantic Web.

You can manually introduce RDFa into existing Web pages. This serves only to enhance the experience ... for machines. As you might imagine, this is an incredibly tedious process, and mechanisms for automation are highly desirable. Some form of templating mechanism, that can be re-used by all pages on your Web site for example, is one viable way for automatically introducing RDFa-based content into static Web pages. In some respects, RDFa is more-easily introduced via dynamically generated Web pages. For example, Ruby-on-Rails provides assistance to developers who seek to introduce RDFa into their Web applications and ultimately their Web pages.

Although RDFa has been in the works since 2004, it only completed the W3C's standardization process last October. With interest from Yahoo! to the Creative Commons, RDFa is being adopted. In addition, an ecosystem of tools and utilities (e.g., parsers, converters, etc.) is developing. Despite such promising signs, RDFa is not the only option for embedding semantic markup in HTML. Embedded RDF (eRDF) and microformats offer alternatives - and in the latter case, uptake has been significant. There is no doubt that mechanisms for embedding semantic markup in HTML are of increasingly signicaificant value. What is debatable, however, is whether RDFa, eRDF or microformats will dominate ...

If you are keen to learn more about RDFa, the following resources will serve as solid starting points:

• Wikipedia's entry on RDFa
• The W3C's RDFa Primer and RDFa Use Cases
• The RDFa Wiki

Even at this early stage, it makes sense to learn more about RDFa, and factor it into your Web site at the earliest of stages.